XORs in the Air: Practical Wireless Network Coding

1
XORs in the Air Practical Wireless Network Coding
Sachin K., Dina K., Wenjun Hu Hariharan R. and
Muriel M.

• ??????? ??
• ?????
• ??M97G0216

2
Outline
3
Introduction

• COPE, a new forwarding architecture that
  substantially improves the throughput of wireless
  networks.
• COPE inserts a coding shim between the IP and MAC
  layers, which identifies coding opportunities and
  benefits from them by forwarding multiple packets
  in a single transmission.

4
(No Transcript)
5

• COPE leads to larger bandwidth savings than are
  apparent from this example.
• COPE exploits the shared nature of the wireless
  medium which, for free, broadcasts each packet in
  a small neighborhood around its path.
• Each node stores the overheard packets for a
  short time.
• When a node transmits a packet, it uses its
  knowledge of what its neighbors have heard to
  perform opportunistic coding.

6

• COPEs design is based on two key principles.
• COPE disposes of the point-to-point abstraction
  and embraces the broadcast nature of the wireless
  channel.
• COPE employs network coding.
• Our work is rooted in the theory of network
  coding, which allows the routers to mix the
  information content in the packets before
  forwarding them.

7

• Allerton 2005 ,this paper departs from our
  previous paper and all prior work on network
  coding in three main ways
• It articulates a full-fledged design that
  integrates seamlessly into the current network
  stack, works with both TCP and UDP flows, and
  runs real applications.
• The implementation is deployed on a 20-node
  wireless testbed, creating the first deployment
  of network coding in a wireless network.

8

• The paper studies the performance of COPE, and
  reveals its interactions with the wireless
  channel, routing, and higher layer protocols.
• Our findings can be summarized as follows
• Network coding does have practical benefits, and
  can substantially improve wireless throughput.
• When the wireless medium is congested and the
  traffic consists of many random UDP flows, COPE
  increases the throughput of our testbed by 3-4x.
• If the traffic does not exercise congestion
  control ,COPEs throughput improvement may
  substantially exceed the expected theoretical
  coding gain.

9

• For a mesh network connected to the Internet via
  an access point, the throughput improvement
  observed with COPE varies depending on the ratio
  between total download and upload traffic
  traversing the access point.
• Hidden terminals create a high collision rate
  that cannot be masked even with the maximum
  number of 802.11 retransmissions.

10
(No Transcript)
11
COPE Overview

• Our protocol is designed for wireless mesh
  networks.
• It uses network coding for unicast traffic.
• The main characteristic of our approach is
  opportunism
• Each node relies on local information to detect
  and exploit coding opportunities whenever they
  arise.
• The scheme has two components
• Opportunistic listening
• Opportunistic coding

12

• Opportunistic Listening
• Wireless is a broadcast medium, creating many
  opportunities for nodes to hear packets even when
  they are not the intended recipient.
• We make all nodes in the network store all
  packets they hear for a limited amount of time T
  .
• Reception reports.

13
(No Transcript)
14

• Opportunistic Coding
• What packets to code, and how?
• Each node should answer this question based on
  local information and without consulting with
  other nodes.
• When the MAC indicates that the node can send,
  the node picks the packet at the head of the
  queue, checks which other packets in the queue
  may be encoded with this packet, XORs those
  packets together, and broadcasts the XOR-ed
  version.

15
(No Transcript)
16

• Learning Neighbor State
• But how does a node know what packets its
  neighbors have?
• Each node announces to its neighbors the packets
  it stores in reception reports.
• A node cannot rely solely on reception reports,
  and may need to guess whether a neighbor has a
  particular packet.

17
Understanding Copes Gains

• How beneficial is COPE?
• Its throughput improvement depends on the
  existence of coding opportunities, which
  themselves depend on the traffic patterns.
• Coding Gain
• We defined the coding gain as the ratio of the
  number of transmissions required by the current
  non-coding approach, to the minimum number of
  transmissions used by COPE to deliver the same
  set of packets.

18
(No Transcript)
19

• What is the theoretical capacity of a wireless
  network that employs COPE?
• The capacity of general network coding for
  unicast traffic is still an open question for
  arbitrary graphs.
• THEOREM 4.1. In the absence of opportunistic
  listening, COPEs maximum coding gain is 2, and
  it is achievable.

20

• CodingMAC Gain
• It turns out that the interaction between coding
  and the MAC produces a beneficial side effect
  that we call the CodingMAC gain.
• The mismatch between the traffic the router
  receives from the edge nodes and its
  MAC-allocated draining rate makes the router a
  bottleneck
• Half the packets transmitted by the edge nodes
  are dropped at the routers queue.

21

• For topologies with a single bottleneck,the
  CodingMAC gain is the ratio of the bottlenecks
  draining rate with COPE to its draining rate
  without COPE.
• what is the maximum CodingMAC gain?
• The maximum possible CodingMAC gains with and
  without opportunistic listening are properties of
  the topology and the flows that exist in a
  network.
• THEOREM 4.2. In the absence of opportunistic
  listening, COPEs maximum CodingMAC gain is 2,
  and it is achievable.

22
(No Transcript)
23
Making It Work

• Packet Coding Algorithm
• To build the coding scheme, we have to make a few
  design decisions.
• We design our coding scheme around the principle
  of never delaying packets.
• COPE gives preference to XOR-ing packets of
  similar lengths, because XOR-ing small packets
  with larger ones reduces bandwidth savings.
• Notice that COPE will never code together packets
  headed to the same nexthop, since the nexthop
  will not be able to decode them.

24

• Searching for appropriate packets to code is
  efficient due to the maintenance of virtual
  queues.
• Another concern is packet reordering.
• We want to ensure that each neighbor to whom a
  packet is headed has a high probability of
  decoding its native packet.
• Let the probability that a nexthop has heard
  packet i be Pi.

25

• Formally, each node maintains the following data
  structures.
• Each node has a FIFO queue of packets to be
  forwarded, which we call the output queue.
• For each neighbor, the node maintains two
  per-neighbor virtual queues, one for small
  packets , and the other for large packets. The
  virtual queues for a neighbor A contain pointers
  to the packets in the output queue whose nexthop
  is A.
• The node keeps a hash table, packet info, that is
  keyed on packet-id. For each packet in the output
  queue, the table indicates the probability of
  each neighbor having that packet.

26
(No Transcript)
27

• Pseudo Broadcast
• This section shows that implementing network
  coding directly on top of 802.11 broadcast may
  produce lower throughput than no coding at all.

28

• 802.11 MAC, which works in two modes
• Unicast
• Broadcast.
• Our solution, called pseudo-broadcast, piggybacks
  on 802.11 unicast, which has a backoff mechanism.
  
• Pseudo-broadcast unicasts packets meant to be
  broadcast

29

• Hop-by-hop ACKs and Retransmissions
• Why hop-by-hop acks?
• Encoded packets require all nexthops to
  acknowledge the receipt of the associated native
  packet for two reasons.
• Encoded packets are headed to multiple nexthops,
  but the sender gets synchronous MAC-layer acks
  only from the nexthop that is set as the link
  layer destination of the packet.
• COPE may optimistically guess that a nexthop has
  enough information to decode an XOR-ed packet,
  when it actually does not.

30

• Asynchronous Acks and Retransmissions
• How should we implement these hop-by-hop acks?
• For non-coded packets, we simply leverage the
  802.11 synchronous acks.
• Unfortunately, extending this synchronous ack
  approach to coded packets is highly inefficient,
  as the overhead incurred from sending each ack in
  its own packet with the necessary IP and WiFi
  headers would be excessive.

31

• Preventing TCP Packet Reordering
• Asynchronous acks can cause packet reordering,
  which may be confused by TCP as a sign of
  congestion.
• COPE has an ordering agent, which ensures that
  TCP packets are delivered in order.

32
Implementation Details

• Packet Format
• Ids of the coded native packets
• The first block records metadata to enable packet
  decoding.
• Expressing asynchronous acks compactly and
  robustly
• To ensure ack delivery with minimum overhead, we
  use cumulative acks.
• Since they implicitly repeat ack information,
  cumulative acks are robust against packet drops.

33

• Reception reports

34

• Control Flow

35
(No Transcript)
36
Experimental Results

• This section uses measurements from a 20-node
  wireless testbed to study both the performance of
  COPE and the interaction of network coding with
  the wireless channel and higher-layer protocols.
• Our experiments reveal the following findings.
• COPE in gadget topologies
• LongLived TCP Flows

37

• UDP Flows

38

• COPE in an Ad Hoc Network
• TCP

39
(No Transcript)
40

• UDP

41
(No Transcript)
42

• COPE in a Mesh Access Network

43
(No Transcript)
44
Discussion and Conclusion

• COPE can be used in multi-hop wireless networks
  that satisfy the following
• Memory
• Omni-directional antenna
• Power requirements
• Our community knows a few fundamental approaches
  that can improve wireless throughput, including
  more accurate congestion control, better routing,
  and efficient MAC protocols.

45
Thank You !

• Add your company slogan